Learning Forum

[Audy Maulizar]

Dear all. I am having a little bit of trouble with setting up a looping transient structural analysis code. For the sake of brevity, I have put only the relevant block of code below. 

!---------------------------------Flutter Analysis Loop---------------------------------
U_fin = 19 !m/s
init_time = 5 !s
final_time = 20 !s
dt = 2.5E-02 !time step in seconds
total_steps = (final_time-init_time)/dt

*DO,i,1, total_steps, 1
    
    /PREP7    
    [Element Modification Code Here]
        
    ALLSEL, ALL
    /SOLU
    time_val = init_time + i*dt
    ANTYPE,,RESTART
    AUTOTS, OFF
    DELTIM, dt
    TIMINT, ON
    OUTRES,ALL,20
    TIME, time_val
    SOLVE
    FINISH

*ENDDO

The general idea behind this looping section is to have ANSYS modify the properties of a structural element in every load step (as shown with the /PREP7 section) then solve the resulting new structural setup (as shown in /SOLU section). The problem I faced was that the DELTIM and TIME commands seem to not receive the values from the parameters "dt" and "time_val" respectively, regardless of how those parameters have been defined beforehand. This resulted in ANSYS returning to the default values for the time step sizes and time at the end of a load step. 

Anyone's help on this problem is very welcome.

[mrife]

Hi Audy Maulizar

The restart database contains the parameters as they are at the first substep of the first load step.  Use parsav and parres to save/resume the parameters within the loop.

Also, I am getting a 'code smell' from the posted script and the "element modification code".  What exactly is going on here?

 

Mike

[Audy Maulizar]

Thank you for your answer Mike. The [Element Modification Code Here] is just a block of code I wrote to change the parameter values for the MATRIX27 elements. I omitted it from this post, because I figured it'd make the code too long for the troubleshooting purpose when it's the looping part that has problems. Here's a more complete version for your reference.

 

!---------------------------------Flutter Analysis Loop---------------------------------
U_fin = 19 !m/s
init_time = 5 !s
final_time = 20 !s
dt = 2.5E-02 !time step in seconds
total_steps = (final_time-init_time)/dt

*DO,i,1, total_steps, 1
    
    /PREP7    
    !Constant Coefficients
    rho = 1.225 ! air density kg/m^3
    B = 11.89 ! deck width
    Le = 15.24! element length
    Pi = 3.14159
    Uh = U_fin*dt/(final_time-init_time)
    init_freq = 0.19788713 ! Hz
    osi_freq_0 = init_freq
        
    ! flutter variables
    Ur = Uh/(osi_freq_0*B)
    omega_h = 2*Pi*osi_freq_0 ! radial natural frequency
    K_h = omega_h*B/Uh
    a_h = 0.5*rho*Uh*Uh*K_h*K_h*Le
    b_h = 0.5*rho*Uh*B*K_h*Le
        
    ! flutter derivatives
    A1 = vala1(Ur)
    A2 = vala2(Ur)
    A3 = vala3(Ur)
    A4 = vala4(Ur)
    A5 = vala5(Ur)
    A6 = vala6(Ur)
    H1 = valh1(Ur)
    H2 = valh2(Ur)
    H3 = valh3(Ur)
    H4 = valh4(Ur)
    H5 = valh5(Ur)
    H6 = valh6(Ur)
    P1 = valp1(Ur)
    P2 = valp2(Ur)
    P3 = valp3(Ur)
    P4 = valp4(Ur)
    P5 = valp5(Ur)
    P6 = valp6(Ur)
        
        
    r,900
    rmodif, 900, 14, -b_h*P5
    rmodif, 900, 15, -b_h*P1
    rmodif, 900, 16, -b_h*B*P2
    rmodif, 900, 26, -b_h*H5
    rmodif, 900, 27, -b_h*H1
    rmodif, 900, 28, -b_h*B*H2
    rmodif, 900, 38, -b_h*B*A5
    rmodif, 900, 39, -b_h*B*A1
    rmodif, 900, 40, -b_h*B*B*A2
    rmodif, 900, 92, -b_h*P5
    rmodif, 900, 93, -b_h*P1
    rmodif, 900, 94, -b_h*B*P2
    rmodif, 900, 104, -b_h*H5
    rmodif, 900, 105, -b_h*H1
    rmodif, 900, 106, -b_h*B*H2
    rmodif, 900, 116, -b_h*B*A5
    rmodif, 900, 117, -b_h*B*A1
    rmodif, 900, 118, -b_h*B*B*A2
        
    CMSEL, S, DampMatElement, ELEM
    EMODIF, ALL, REAL, 900
        
    r,901
    rmodif, 901, 14, -a_h*P6
    rmodif, 901, 15, -a_h*P4
    rmodif, 901, 16, -a_h*B*P3
    rmodif, 901, 26, -a_h*H6
    rmodif, 901, 27, -a_h*H4
    rmodif, 901, 28, -a_h*B*H3
    rmodif, 901, 38, -a_h*B*A6
    rmodif, 901, 39, -a_h*B*A4
    rmodif, 901, 40, -a_h*B*B*A3
    rmodif, 901, 92, -a_h*P6
    rmodif, 901, 93, -a_h*P4
    rmodif, 901, 94, -a_h*B*P3
    rmodif, 901, 104, -a_h*H6
    rmodif, 901, 105, -a_h*H4
    rmodif, 901, 106, -a_h*B*H3
    rmodif, 901, 116, -a_h*B*A6
    rmodif, 901, 117, -a_h*B*A4
    rmodif, 901, 118, -a_h*B*B*A3
        
    CMSEL, S, StiffMatElement, ELEM
    EMODIF, ALL, REAL, 901
        
    ALLSEL, ALL
    /SOLU
    time_val = init_time + i*dt
    ANTYPE,,RESTART
    AUTOTS, OFF
    DELTIM, dt
    TIMINT, ON
    OUTRES,ALL,20
    TIME, time_val
    SOLVE
    FINISH

*ENDDO

[dlooman]

It's possible you could change the matrix27 values in solution simplifying the analysis.  Worth a try anyway.

[mrife]

Hi Audy Maulizar

Dave Looman is right, defining a real constant can be done in the solution processor (as well as in the preprocessor).  And I don't see a reason in the script to have the emodif commands, so you don't need to exit out of the /solu processor.

[Audy Maulizar]

By not having emodif command, I assume the RMODIF commands are already enough? 

[mrife]

Hi Audy

By the script shown yes assuming that say the elements of component DampMatElement aleady were assigned the real constant (ID) 900.